CN1805447B - Method for assigning a priority to a data transfer in a network, and network node using the method - Google Patents

Abstract

A data transfer in a network comprises a first node sending out a request for a particular data unit, a second node receiving and analysing the request, detecting that it may provide the requested data unit and sending to the first node a message indicating that it may provide the requested data unit, the first node receiving and selecting the message and sending a second request to the second node to request transfer of the particular data unit, and the second node transferring the particular data unit upon reception of the second request. A method for assigning a priority to such data transfer in a network comprises the first node assigning an identifier corresponding to a first priority to the request, the second node evaluating the identifier and, based on the identifier, calculating a second priority and assigning the calculated second priority to said transfer.

Description

For the data in the network transmit method of distributing priority and the network node that uses this method

Technical field

The present invention relates generally to network service.Particularly, the present invention relates to a kind of method and a kind of network node that uses described method that distributes priority that transmit for the data in the network.

Background technology

In the network such as distributed memory system (DSS), data can be transmitted will carrying out of the task that is interpreted as.Data transmit usually in response to request or task.For example, task can be that search mission or data transmit task, between the related node the characteristic information stream takes place in task.Usually, several (data transmission) task may take place simultaneously concurrently.Because the limited capacity aspect bandwidth, memory space or other parameters, this will lead to a conflict or the bottleneck situation.

The resource that can attempt to distribute another node based on the different nodes in the network (for example, the owner zone described in the European patent application EP 1 427 141) of equity is as memory space or transfer rate etc.If available resources are not enough to manage whole requests, can find clever method to avoid this bottleneck or conflict.This should finish automatically, does not promptly need user interactions.But, in some cases, if user or application program can revise the solution of finding out automatically may be better.

Conflict and bottleneck management mean based on the communication between a plurality of control messages nodes.These control messages also are the parts of language, for example distributed storage communication and control language.

Summary of the invention

The invention provides and manage these conflict and possibilities of bottleneck automatically, and be the device that user or application program are provided for revising the result of automatic acquisition simultaneously.It comprises so-called recessive priority of ground floor and the so-called dominance priority of the second layer based on the definition of double-deck priority system, and wherein recessive priority is arranged dominance priority usually.Therefore, only under the identical situation of the recessive priority of a plurality of tasks, just use dominance priority.In two layers each can be subdivided into different grades.

Advantageously, the present invention only needs communication work considerably less in the network.In addition, can improve the data throughout in the network, utilize memory capacity better, and the availability of improving data.

According to the present invention,, manage or avoided conflict and bottleneck at aspects such as memory space, transfer rate, node availabilities by using one group of priority and the rule of using by the node in the network.Rule is that node is intrinsic, comes calculating priority level according to two steps, as double-deck priority.Ground floor is so-called recessive priority, defines all to relate to rule or the relation that node all meets.Second layer priority is called as dominance priority, by user or application program definition.

Two-stage priority notion has the following advantages: its use task and/or node implicit precedence level, be herein referred to as " recessive priority " and need be by the definition of user or application program, Fu Jia dominance priority relates to the distribution of priority simultaneously, as can be by exchange of user or application program and the information that changes.In other words, can produce recessive priority automatically, and need not user's input.User or application program can be carried out the distribution or the change of dominance priority when thinking fit.

The invention has the advantages that: can suitably manage or avoid conflict and bottleneck, for example in the DSS that is embodied as owner zone, therefore, improved data throughout, utilized memory capacity better, improved availability of data, and prevented network congestion.

The method according to this invention is a kind of for the data in the network transmit the method for distributing priority, and described data transmit and comprise: first node, first request of sending the data cell of specific data cell of indication or particular type; At least Section Point receives and analyzes described first request, and described Section Point detects it can provide the data cell of being asked, and it can provide first message of the data cell of being asked to send indication to described first node; Described first node receives and selects described first message, and sends second request to described Section Point, and request transmits described specific data cell; And described Section Point transmits described specific data cell receiving described second when request.Described method comprises: in first step, described first node is described first request or described second request or these two request distribution markers, and described identifier is corresponding to first priority; In second step, the described Section Point assessment described identifier corresponding with described first priority according to described identifier, calculates second priority; And in third step, described Section Point transmits described data cell specific, that asked, wherein gives described transmission with second priority assignment that is calculated.It should be noted that: the transmission of the data cell of being asked needn't be pointed to the first node of the request of sending.Same possible be: the 3rd node is the receiver of the data cell that transmitted, and first node is just initiated node, for example, because it has user interface, schedule management device etc.In this case, first node will at least the second request also to send to described the 3rd node be useful.

Corresponding equipment comprises the device that is used to carry out each method step.

For example, the data cell of above-mentioned specific data cell or particular type can be have specified title film video data or wherein comprise the video data etc. of whole available movies of specific actors.This information can be associated with data cell, for example, as metadata token, and can be XML form for example.

Advantageous embodiment of the present invention is disclosed in dependent claims, the following description and drawings.

Description of drawings

With reference to the accompanying drawings, exemplary embodiments of the present invention is described, wherein:

Fig. 1 is two real-time streamed situations with enough bandwidth;

Fig. 2 is that two streams with not enough bandwidth transmit;

Fig. 3 has the situation that real-time stream transmits and the while file transmits;

Fig. 4 has the situation that two files transmit, and wherein a dominance priority that transmits task is made amendment;

Fig. 5 is that two files transmit, and second first transmission that is transmitted in of being asked begins before;

Fig. 6 is that two files transmit, and wherein the latter has inherited priority from search mission;

Fig. 7 is the flow chart of the inventive method; And

Fig. 8 be under the situation of capacity limit, the illustrative case of carrying out content replication.

Embodiment

With the owner zone based on peer network architecture is example, present invention is described, wherein node has independent node identifier and shared peer-group identifier, and the node that belongs to peer-group can freely communicate each other, exchanges messages and other data etc.Also can be applied to the network of other types, particularly favourable for the network that its node is organized quite autonomously.

1, priority notion

The present invention has introduced the thought of two-stage notion, relates to the difference between ground floor and the second layer priority: ground floor or recessive priority are the relative priority levels, or the included priority relationship that node met, for example, and the peers include both in the owner zone.It does not have dominance numerical value associated therewith, for example numerical priority or numeral.Therefore, the inherence " knowledge " of recessive priority set expression node promptly depends on one group of rule that it met.Advantageously, can automatically produce recessive priority, thereby user or application program needn't define to it.The second layer or dominance priority relate to the distribution of priority, and for example digital or other identifiers are as the information that can revise or remove.If think fit, user or application program can distribute or revise.Dominance priority can be relative, for example " height " and " low ", or integer, or more generally, be any order item.Give task with the dominance priority assignment of task, and if desired, when for example in determining two tasks which has higher priority for hardware access, memory space, disposal ability etc., the dominance priority of itself and another task can be compared, to make decision.

1.1, recessive priority

Node is embodied as meets following recessive priority rule or relation,, and avoid node in the owner zone and conflict and the bottleneck between the action thereof so that help to manage smoothly transmission.

Basic rule is: " first come, first serve ".For example, by assessing by setting up task and setting up the defined TaskInitTime parameter of node of the time started of task and realize.For example, task can be that search mission or data transmit task, and has the characteristic information stream that takes place between the node related in task.Its whole actions of each node administration in the owner zone, that is, early initiating of task has the priority of the task of being higher than later initiation.Early the message that receives has the priority that is higher than the later message that receives usually.This means that node responds the request that it receives according to the order (being provided by the TaskInitTime parameter) that request is initiated usually.Therefore, it is helpful having shared time reference in related whole nodes.

A scheme of the present invention is: as the exception of this rule, data transmit task can inherit its priority from associated search mission formerly to a certain extent.This is useful, because transmit task and initiate search mission for the data of setting up at the content that is found usually.Therefore, the content that node is guaranteed with early searching request is relevant transmits, allowance time period Twft after the TaskInitTime of searching request (" wait for transmit " time, for example 5 seconds) has the priority that is higher than with in the relevant content transmission of back searching request.But other tasks still can have higher priority, and for example, under the situation that begins immediately to transmit at needs, for example, for the task of record situ flow, node can be treated this special case especially.

As second rule, consider the every other transmission of carrying out or arranging that relates to each node, but, just allow beginning task or data to transmit only in the required resource time spent.This means that node at first checked the resource of the node will be used for task before the initiation task, perhaps may be the resource of all nodes in the owner zone, to obtain overview.At this particular moment, it initiates to transmit, and includes only and have enough memory capacity and those available nodes of transmission capacity (i.e. speed that may transmit and quantity) at that time.This expression source and destination ground node.If desired, the transmission that the node delay will be carried out is till can transmitting after a while.Related node distributes corresponding resource in the transmission.For example, it can cancel distribution by the cancellation task.Therefore, thus prevented the situations that two tasks are blocked each other and blocked whole network.

As three sigma rule, should not interrupt ongoing transmission, unless being initiated node, it cancels clearly.This means that node cannot cancel from transmission other nodes, that moving, set up its oneself transmission to obtain resource.The node that only allows to initiate to transmit is cancelled this transmission.Then, if desired, it can set up another transmission.

As the 4th rule, consider the every other transmission of carrying out or arranging that relates to each node, but only be about to the time spent that just permission is arranged transmission, promptly in ongoing transmission or after being near completion in its resource that takies.This means that node was at first checked in particular moment will be used for the availability of resource that data transmit the node of task.Only constantly for these nodes and this, on the node of destination sufficient memory capacity can with and the source and destination ground node on sufficient transmission capacity (i.e. speed of Chuan Songing and quantity) but the time spent, just initiation transmission.Then, related node distributes resource separately at transmitting the moment that should take place.Can cancel distribution at any time by the cancel transmission task, no matter transmit and whether begin.Therefore, each node that its resource can be offered other nodes have timetable when be used for control, by which node or purpose " reservation " resource why.

As the 5th rule, real-time or stream transmits has the priority that is higher than non real-time or file transmission.Under more general situation, real time data is the data that can not reduce its source rate under the prerequisite of reproducing quality not reducing.Thought is: file transmits and can carry out with any bit rate usually, and cross over any duration feasible according to Internet resources, and in real time or stream transmit (for example audio frequency and/or video data in real time or stream transmit) need carry out with accurate timing, and may relate to and reproduce the needs that the user consumes the content of (as watch or listen to).Node can for example use specific request message by changing bit rate and transmitting the duration, as " ModifyTransferRequest (' modifications ') " etc., slows down or quickens ongoing non real-time/file transmission.The product of transfer rate and transmission duration is file size and thereby remains unchanged.The node of initiation task forbids that a kind of possibility of doing like this is to introduce the task relevant parameter, and as AllowTransferSpeedChange, and it is set to "No".

The 6th rule is that the transmission that is used to write down always has the priority that is higher than the transmission that is used for playback.This rule is subordinated to last rule, and promptly file transmits always to have and is lower than streamed priority.Suppose to exist the time restriction of recorded content,, and also can carry out after a while the playback of content because it may be now available but unavailable after a while.Therefore, if logger task and the competition of playback task, logger task is given with resource allocation in node select ground.Itself in addition can cancel the playback task and realize logger task.This can occur in application program or user class, perhaps takes place automatically, if application program or user allow usually.For example, if for particular moment, arranged the playback transmission, and application program wishes to write down another content in the identical moment, but resource do not allow, then application program can be cancelled the playback of having arranged and transmits, and instead arranges new record to transmit.

For example, in the home network with two recording equipments, a playback apparatus, a receiver and a display device this situation can take place.As user during at the film of watching on the display device from the playback apparatus playback, one of recording equipment is writing down the video flowing from receiver.Suppose after a period of time, the memory of recording equipment is full of, and suppose that network and recording equipment can seamlessly continue record on second recording equipment, then the traffic carrying capacity on the network may be higher during switching to second recording equipment from first recording equipment.But this extra traffic carrying capacity is essential for record, therefore has the priority that is higher than playback of data.In this case, can accept interrupt playback in short time, so that make the record data unanimity.

1.2, dominance priority

Except above-mentioned recessive relatively priority, the present invention also uses optional dominance priority, as " low " and " height " or integer or more generally, and based on any order item of the dominance Priority parameter that can be associated with task.For example, can alternatively dominance Priority parameter be distributed to task by the node of the task of initiation or by the user.Can think that also using dominance priority is the thing of application program.Node can send a request message by each related in the item task other nodes (for example, " ModifyTransferRequest (' modification ') "), revises the Priority parameter, thus and the dominance priority of the task of modification.

Under any circumstance, recessiveness or ground floor priority are all arranged dominance priority.Therefore, only when a plurality of tasks have identical recessive priority, just use dominance priority.If the equipment each run is more than a task, then it sorts to these tasks according to the recessive priority of task, and under the situation of identical recessive priority, if distributed dominance priority, then sort, and sort according to this its resource is provided according to its dominance priority.

If associated user or the application program that allows on this node provide correct use key to it, then only to allow node to revise not to be dominance priority by the task of its initiation.This is the parameter that is associated with each content, and is defined at this purpose by the user alternatively, and for example, can organize relevant with the user of specific care.Can also there be the node of the application program of initiation task to revise dominance priority by operation, perhaps in one embodiment, revises by any node in the owner zone.In this case, anyone in the owner zone can revise the dominance priority of any task with related UseKey parameter.

Below be the example that has defined two dominance priority " low " and " height ", but also can be applied to any precedence scheme.If for task A, undefined dominance priority, this can use following rule for its undefined (or default) value:

Have dominance priority " height " if having another task B (may be the competition task) of identical recessive priority, think that then undefined (or default) dominance priority of task A is " low ";

Have dominance priority " low " if having another task B (may be the competition task) of identical recessive priority, think that then undefined (or default) dominance priority of task A is " height ".

Only this means when want with it to give this task with dominance or the second level " height " priority assignment when treating even more importantly than other tasks with identical recessive priority, then opposite for " low " priority.

If possible, must implement to have the recessiveness that is higher than other tasks or the task of dominance priority, so that its demand is satisfied than other tasks better at aspects such as memory capacity, transfer rates.Should after handling above-mentioned higher priority task, implement to be set to the task of low dominance priority with residual capacity.

1.3, the enforcement of priority rule

In order to implement above-mentioned priority rule, each node can will be referred to it whole carrying out and/or arrangement task is stored in " task and schedule database ".Store these tasks according to the time sequencing of initiating (according to its TaskInitTime), and identify with its task identifier TaskID separately.When finishing, removal task from database.Each node is used above-mentioned priority dependency rule when initiating or serve request.

But Fig. 1 shows two and implements stream and transmit the situation that Tr1, Tr2 have identical recessiveness and dominance priority and enough bandwidth B time spent.At t _TRQ1, request first transmits Tr1, and is to t _SRQ1The response of searching request.But the delivery time interval T is waited in its definition after request _Wft1Beginning is so that check another transmission of whether having asked to have higher priority.In Fig. 1, be not this situation, so at t _SRQ1+ T _Wft1, first transmits Tr1 begins.When the first transmission Tr1 is carrying out, t _SRQ2Second searching request cause t _TRQ2Second request of transmitting.Second transmits Tr2 can be at t _SRQ2+ T _Wft2Beginning is because available data rate or bandwidth B _MaxThe summation R that is higher than the desired data rate ₁+ R ₂t _TRQ1The transmission request also can be later than t _SRQ1Searching request after T _Wft1

Fig. 2 shows the T of second searching request after first searching request _Wft1Interior moment t _SRQ2The situation that arrives.In addition, for example, if recessive priority equates that then owing to dominance priority, second priority P 2 that transmits Tr2 is higher than the priority P 1 of the first transmission Tr1.But, do not have enough bandwidth to can be used for carrying out simultaneously these two transmission.Therefore, because t _SRQ2＜t _SRQ1+ T _Wft1, second transmits Tr2 at first begins, and another transmission Tr1 t after finishing Tr2 of request early _E2Beginning.This is the exception of previously mentioned first come, first serve rule shown in Figure 1.If in Fig. 2, second searching request a little later arrives again, i.e. t _SRQ2＞t _SRQ1+ T _Wft1If, then have identical recessive priority, for example, be real-time stream and transmit, first transmits Tr1 begins.

It is later that Fig. 3 shows second searching request, i.e. t _SRQ2＞t _SRQ1+ T _Wft1Thereby first transmits the situation that Tr1 has begun.But second searching request has higher priority, and for example, Tr1 is that file transmits, and Tr2 is that stream transmits in real time, and available bandwidth B _MaxBe not enough to carry out concurrently these two transmission: B _Max＜R ₁+ R ₂In this case, in any case because its higher priority, second transmits Tr2 at t _SRQ2+ T _Wft2Beginning, and the ongoing first transmission Tr1 has to the data transfer rate R of reduction when Tr2 carries out _1red: B _Max＞R _1red+ R ₂Less bandwidth left B _Max-R _1red-R ₂Still available, so that the communication information in the realization network.At t _E2, finish second and transmit after the Tr2, first transmits Tr1 obtains its whole bandwidth R once more ₁Effect is: first transmits Tr1 needs the more longer time, and flow data transmission Tr2 can carry out in real time.During Tr2, the bit rate adjustment of Tr1 is not influenced the quality of data, because Tr1 is a real time data.Advantageously, these two transmission can not blocked each other, and still leave the bandwidth capacity that is used for network service.

Fig. 4 shows the situation of using dominance priority.At t _SRQ1, in home network, initiate first searching request, cause at t _SRQ1+ T _Wft1Beginning, have first a recessive priority P ₁First transmit Tr1.After a while, at t _SRQ2, second searching request causes at t _SRQ2+ T _Wft2Beginning, have a P of equaling ₁The second recessive priority P ₂Second transmit Tr2.Be the file transmission and have identical dominance priority, be low, undefined or high.Therefore, two transmission obtain identical data transfer rate: R ₁=R ₂But, after a while, at t _U, the user determines to change the priority of the first transmission Tr1, for example, writes the removable dish that access customer is wanted very fast use because will transmit Tr1.Therefore, the user can become higher with the first dominance priority that transmits Tr1, as shown in Figure 4, perhaps the second dominance priority that transmits Tr2 is become lower.As a result, first be transmitted in t _UObtained bigger data transfer rate afterwards, and at t _E1, finish very soon.Afterwards, second transmits Tr2 can obtain bigger data transfer rate, thereby in situation shown in Figure 4, two to transmit required total time identical.

Although only, exemplarily show described fundamental mechanism, can use it for the transmission of any amount, and can make up it at two transmission.For example, in Fig. 4, can be at t _E1T afterwards, _E2Before, utilize mechanism shown in Figure 3, ask and begin to have another transmission of higher priority.

Fig. 5 shows another embodiment of the present invention.Each interval is sent first request RQ1 that transmits at file and the second request RQ2 that transmits at file shortly.Its priority P can be understood as from default value P ₀Beginning is risen continuously, realizes the first come, first serve rule thus.Comparatively fast replied the second request RQ2, and corresponding transmission T2 begins (wait delivery time section T that may be after replying at TS2 _WftAfterwards), and do not find relevant content yet, for example, hurry because have the node of this content in the first request RQ1.The priority P of ongoing transmission T2 ₂Keep constant, and the priority of first request rises further, up to replying this request, and till transmitting T1 and beginning.Priority remains on the numerical value that TS1 is had when transmitting beginning.Because first to transmit the priority of T1 higher, and these two transmitted T1, T2 and be the non real-time file and transmit, and in the present embodiment, first transmits the bandwidth that T1 obtains to transmit more than another T2.Therefore, can finish at TE1 quickly, should do like this, because earlier it is asked.

Figure 6 illustrates similar situation.But here, the second request RQ2 ' has the priority that is higher than the first request RQ1 '.For example, the user gives this request RQ2 ' higher dominance priority.All ask the non real-time file to transmit.When replying for second when request, transmit from request RQ2 ' and inherit its priority P 2 ', and can (may be at T at TS2 ' _WftBeginning afterwards).When replying the first request RQ1, it has the priority P 1 ' that is lower than second transmission, and therefore has to less bandwidth resources, transmits T2 ' up to second and finishes.

2, conflict and bottleneck and management thereof and avoid scheme

Two or more operations compete each other and situation about repelling under, clash, thereby can not carry out all operations.For example, first application program may attempt to delete this content when Another Application program reading of content.Therefore, term " conflict " is represented as the system's conflict in the network systems such as DSS, and has been described the situation that can not carry out required task.But existence overcomes the mode of conflict.As the possibility in the above-mentioned example, can after reading task, carry out the deletion task, perhaps the task of reading can be cancelled, thereby the deletion task can be carried out.

Bottleneck is a physical constraint, for example low throughput rate or memory capacity, high latency etc.Therefore, be to handling or the limiting factor of task.Thereby in this application, " bottleneck " expression can be carried out desired task but just have the situation that limits.Different with conflict, bottleneck can't block or the prevention task.

With the lower part a plurality of conflicts and bottleneck and management thereof have been described.And the scheme of avoiding to it proposed.

2.1, the conflict and the management

For example, can clash in the following areas:

Memory capacity: for example, transmit for data, the memory capacity of destination node may be not enough;

Transfer rate: for example, transmit for data, the available transfer rate of source or destination node may be not enough;

Transmit quantity: may reach the transmission quantity that node can be managed, transmit request thereby can not handle other;

Visit: two nodes may attempt to visit simultaneously the resource (for example, memory capacity, transfer rate, disposal ability etc.) of the 3rd node;

There is not response: for example,, can receive desired nothing response because node does not connect;

Application program or user's rights and interests: user/application program can not be visited required content, because Usekey the unknown associated therewith;

Formerly distribute: when node during to another node request specific resources, may receive sufficient resources can with reply; But when it attempts Resources allocation, may be rejected, because the 3rd node has distributed this resource simultaneously;

The node availability: as long as node is unavailable in network, for example, owing to disconnect or temporary transient outage, its resource (as storage content thereon) is unavailable for other nodes; Even node may become unavailable before the transmission that transmits when carrying out or arranging begins.

Can use message and control metadata to overcome conflict on the memory capacity.For example, in order to overcome storage space conflict, application program or user can determine deletion or move and comparatively lose interest in or comparatively unessential content.For example, this can be according to user's preference decision.Therefore, reserved the space that is used for new record.In order to overcome the conflict on the transfer rate, can reportedly send by the order actual figure.

Can be used as the precautionary measures and carry out resource management continuously, perhaps only in case of emergency carry out.In case node receives or sends corresponding request, just the resource in the node is distributed, for example, will be used for content and transmit.In this stage, searching request does not hint the distribution of resource yet, because usually and do not know the intention and the decision of user or application program; For example, may find several occurrences, then must make a choice.But data transmit and may follow thereafter.Therefore, the objective of the invention is: searching request early causes the higher priority of the transmission of Search Results.This will further explain in detail following carrying out at the part of priority.With the initiation time of other node communication search missions related in the task, the i.e. time of definition of T askID.

In order to improve availability, can be on two or more nodes with important content replication and redundant storage.Therefore, can be stored in content on the present disabled specific node to another node visit.This is the problem of application layer or intermediate controlled layer.For example, system can know or inquire that the user in owner zone is interested in those classifications, and creates the copy of corresponding contents automatically.The known content that will be recorded on the removable medium also can be duplicated by system, and it is stored on the mounting medium that can be used for the owner zone.Therefore, software need be followed the tracks of the available moment of node, and the user that what is thought is important.

If identical content can obtain on redundant ground, then also can be used to overcome specific visit or transfer rate conflict on different nodes.For example, if two nodes attempt to visit the identical content on the 3rd node, one of them can be redirected to the equivalents on another node.If node is found identical content on different nodes, it can select to provide the node of high transfer rate.

If in officely do honest work when carrying out and the node of the source of non task or destination becomes unavailable, this is not problem usually.

Become unavailable if initiate the node of searching request, related other nodes are regarded this disappearance as to task cancellation in the search mission, and from its task memory deletion task and parameter thereof.

If request provides the node of relevant with content or relevant with its place capacity information to become unavailable, then just do not respond.Requesting node is accepted this situation after overtime.

If initiate that content transmits but be not that the node of source or destination itself becomes unavailable, then just with the beginning of transmission with finish relevant notification message and can not arrive.After successfully transmitting, the source and destination ground node is this task of deletion and parameter thereof from its task memory as usual.When but the node of initiating task becomes the time spent when carrying out once more transmitting, some notification messages will arrive, and task is almost finished as usual.But become the time spent once more after the basic point of the task of initiation is transmitting, it is analyzed TaskInitTime the current time is added that (expection) transmits the duration, and deletes this task and parameter thereof from its task memory; It can transmit content by search on the node of destination and check and whether successfully finished transmission, and whether decision needs to attempt this transmission once more by initiating newly to send.

If source or destination node become unavailable in transmission when carrying out, for example because transmission can not be successfully finished in outage or disconnection.Usually, but some exceptions are arranged, related node will think that task is cancelled, and delete this task and parameter thereof as quickly as possible from its task memory.There are different situations and possibility:

If it is unavailable that source node becomes during transmitting, then the destination node can (a) be deleted the content that it has received; Perhaps (b) keeps this content, and for it distributes new ContentID, and mark concluding time or end bit; Perhaps (c) keeps this content, keeps original ContentID, and mark concluding time or end bit, recovers to transmit so that attempt after a while.Then, it will transmit task flagging for interrupting in its task memory.If it is not source or destination node itself, it should give interrupt notification the node of this task of initiation.Can use particular message such as (" interruptions "), and wait for cancellation request a little from other nodes as TransferStatusInformation.In situation (a) with (b), the node of the destination node and the task of initiation is this task of deletion and parameter thereof from its task memory then; But source node carries out same operation in the time spent that becomes once more.In situation (c), source node should be attempted to get in touch in the destination, in case that it becomes once more is available, recover to transmit from interrupted point, and the node (using as TransferStatusInformation message such as (" recoveries ")) of notice initiation task; If at given time period T _WuaIn (" waiting by the time available " time, for example a week), it is available that source node does not become yet, and then the node of the destination node and the task of initiation moves according to situation (b).

If it is unavailable that the destination node becomes during transmitting, then source node stops to send data, interrupt notification is given the node (unless it is source or destination node itself) of initiation task, for example, use TransferStatusInformation (" interruption ") message, and wait for cancellation request a little from it.Then, this task of deletion and parameter thereof from its task memory.According to which resource can use, initiate the node (being not source or destination node itself) of task and can (a) attempt to initiate the transmission of corresponding contents to another destination node; Perhaps (b) waits for always, becomes available once more up to last destination node; Under a kind of in the back situation, task and parameter thereof are kept in its task memory, and transmission is labeled as interruption.If last destination node becomes available once more, it checks its task memory, and detect interruption, trace into and transmit interrupted point, and transmit data from this point by the request source node, recover from this point to transmit, and utilize, the node of notice initiation task as TransferStatusInformation (" recovery ") message.Initiate node and can delete the transmission task in situation (a), the result is that the destination node should be deleted the content that has transmitted, and perhaps in situation (b), with the same action the during the common transmission, i.e. notice is finished in wait task.

Also can arrange to transmit at particular moment.If node is unavailable when safe transmission should begin, following situation is possible:

If source node is unavailable in the zero hour of arranging to transmit, then the destination node uses as TransferStatusInformation message such as (" beginnings "), this event notice is given initiated node (if it is not source or destination node itself).Then, it is waited for a little from the cancellation request of initiating node.If do not receive the cancellation request, it is at given time period T _WuaAttempt once more in (for example one hour or a week) beginning to transmit.During this period, initiate node and can cancel this task at any time.Under the situation of cancellation, perhaps at time period T _WuaDuring end, destination node and initiation node are deleted this task and parameter thereof from its task memory.But source node carries out same operation in its time spent that becomes once more.If at T _WuaInterior source node is available once more, and can successfully begin to transmit, negligible delay then, and use common message flow.

If the destination node is unavailable in the zero hour of arranging to transmit, it can not begin the request source node in the moment of having arranged and transmit content to it.Source node is given this event notice and is initiated node (if it is not source or destination node itself), for example, uses TransferStatusInformation (" beginning ") message.

According to available resources, initiate node can (a) to wait for that the destination node becomes once more available, begin then to transmit; Perhaps (b) sends the cancellation request.In situation (b), can select another destination node.In situation (a), source node and initiation node are kept at time period T given in its task memory with this task and parameter thereof _Wua, and after this it is deleted.Time spent is carried out same operation once more but the destination node is at it.If the destination node is at T _WuaIn available once more, its request source node for data forwarding.If can successfully begin to transmit, use common message flow.If now source node is unavailable, the destination node should move like that according to the above-mentioned source node disabled situation that becomes.

Under any circumstance, any node should be deleted from its task memory and incur loss through delay the time T that exceeds defined _WuaAny task, comprise its relevant parameter.

2.2 bottleneck and management thereof

For example, bottleneck can take place in the following areas:

Memory capacity: for example, the memory capacity of destination node may be not enough to reportedly send according to the such actual figure of ask;

Transfer rate: for example, the available transfer rate (bandwidth) of source or destination node may be not enough to reportedly send according to the such actual figure of ask;

Disposal ability/time: for example, memory node can not simultaneously or be carried out the searching request that all receives in the expected time.

Can use message and control metadata to overcome bottleneck on memory capacity and/or the transfer rate.In order to overcome the bottleneck on the transfer rate, application program or user can determine to transmit with lower bit rate, in non real-time content, and no matter it is to flow content or non real-time file content in real time, so the delivery time will be longer.In case it is available that resource becomes once more, can increase bit rate once more, and shorten the delivery time.As required, can use the device that is used to adjust the bit rate that file transmits.

When search in real time the stream content so that send portable with low transfer rate it or during mobile device, can in searching request, comprise Maximum Bit Rate.Only preserve required content and will reply this request with the equipment of bit rate coupling.If under the situation of disposal ability/time-related bottleneck, memory node can not simultaneously or be carried out the searching request that all receives in the expected time, and it will periodically be communicated by letter, and it is still being searched for.In a word, it can manage whole searching request, if necessary, carries out sequence management.

Also there is main possibility on application layer, exceeded the scope of message and control metadata in fact, can be used for overcoming bottleneck.For example, under the situation of the bottleneck aspect transfer rate or the memory capacity,, can carry out the expection that is used for playback or record purpose according to the bit rate that reduces (the therefore quality that reduces) and flow transmission in real time if node has such ability.

2.3, for avoiding conflict and bottleneck

Do not need always to be in and clash or the situation of bottleneck.For example, can take following steps in advance, to avoid or to reduce the number of times of bottleneck and conflict.

Keep the transmission capacity to use: in order to use for arbitrary node leaves some transmission capacity at any time, should arrange the transmission in the owner zone (especially when regarding monolithic as), can be with (being that MaxStreams-ActiveStreams is at least 1) thereby each node is had at least at the capacity of a transmission.Initiate node and need consider this point.Usually, in order always to visit the content that is stored on the node, if possible, the last available transmission of node should be kept to playback.When having record request and having only a node to use, only leave one not with the node time spent of transmitting but perhaps have only, then any one in this node or these nodes should be replied this request and be write down this content; In every other situation, each node should will not given playback with transmitting to keep at last.

But, should be noted that and arrange to transmit that for example having arranged to transmit simultaneously, the whole of distribution node may transmit (MaxStreams).

Keep memory capacity to use: in order to use for arbitrary node leaves some memory capacity at any time, should analyzing stored on node the content in (or in whole owner zone), can to application/user provide repetition or similarly content, with such as the visit seldom or the content of other criterions couplings such as visit, so that deletion.Perhaps, can notify and ask the user to obtain more storage capacity.

When arranging record request, can analyzing stored on node or the content in the owner zone, and can notify user or application program whether to store same or analogous content.Described analysis should be considered that the content of having stored is whether complete and be enough quality.Then, application program can advise not carrying out new record, perhaps deletes other versions, if for example it has than low quality or imperfect.

Early warning: it is not reduced to one node other nodes around can be in the owner zone with transmission quantity and sends DeviceCapabilitiesInformation message.

Soft disconnection:, can notify it to be about to message to other nodes when " soft " disconnects rather than " firmly " disconnected node when being possible.For example, this can close whole application programs by adopting at application-level, perhaps a class of being sent by the user closing/disconnecting to move to wait and realize based on software.

Below be the simple case of having described the application of the present invention in distributed memory system, the control language that is used for distributed storage management comprises association messages and control metadata.Different message or task with as message parameter or independent variable and the specific control metadata that is included in wherein use.For convenience of description, represent message with message name and independent variable, for example:

DeviceCapabilitiesInformation(

Sender，Receiver，TaskID，DeviceType，DeviceServices，

MaxBitRate，FreeStorageCapacity，…)

Although each message has the MessageID of himself,, omitted MessageID for easy.Described situation is based on the example network at distributed storage shown in Figure 8 (owner zone).Described network is by fixed memory device or node S0 ... S3 (as PDR, HDD, CD) and portable memory apparatus or node P constitute.Each node P, S0 ... S3 can run application and have user interface or remote controller (equipment/node that also can regard separation as).Can be tuner/receiver device (as DVB-S or DVB-C), AV demonstration/output equipment, be used for the ADSL Modem that inserts the internet or gateway etc. for may expanding of home network.In illustrative case, a node S0 is generally used for carrying out alternately with distributed memory system.In this case, the user wants to use reproducting content under the situation of the memory capacity in the network in capacity limit and balance.At first, by node S0 ... network startup that S3, P constitute and operation do not have content to transmit all node free time take place.On content replication any one in fixed memory device S1, S2, the S3 that the user wants to be stored on the P.Content replication is not used on the permanent plant of memory capacity to providing at most.

The user utilizes equipment S0 to search for required content: all devices of equipment S0 in network sends search request message.Equipment P receives this message, detects it and comprises this content, and reply S0.But, in the variant of this situation, can use equipment P to replace S0 to initiate to search for task with reproducting content.In this case, node P can be to the answer of the content that himself sends relevant and request coupling, and the corresponding information of acquisition from its content data base just.

Because the user wants content stores to use equipment S0 to inquire its storage and transmission capacity to equipment S1, S2 and S3 on any fixed memory device.S1, S2 and S3 are notified to S0 with its place capacity, and promptly it all has sufficient can not using with transfer rate.For equipment S1, observe not using the restriction of memory capacity, and S3 provides maximum free capacities.Therefore, equipment S0 request P sends content to S3, thereby uses memory capacity available in the network in the mode of balance.After finishing the associated data transmission, P utilizes message to notify S3.After recorded content, S3 completes successfully to the S0 notice.

For example, the balance of memory capacity is used in the network, i.e. managed storage space between node, and can mean is providing the highest not with transfer rate or the highest definitely or do not use on the node of memory capacity relatively with content record, and situation is similar therewith.Memory device in the network can be considered to " monolithic ", and the user need not it is distinguished.But it is a kind of possibility mode of the memory capacity in the supervising the network that the balance of memory capacity is used.When reproducting content, for example, under the situation of capacity limit, also can use other strategies.

Following typical message sequence takes place in this case: all message all comprise transmit leg and recipient's identifier and the peculiar parameter of each type of message.

Suppose that the user wants to search for the content of specific or particular type, for example title is the film of " Octopussy ".As the result of input, S0 equipment sends following searching request to all devices; Because S0 has the understanding in advance of some relevant S2, or the S2 that cherishes a special interest, S0 sends following message to S2 particularly:

ContentInfoRequest(

Sender＝NodeID(S0)，Receiver＝all，Receiver＝NodeID(S2)，

TaskID＝abc，TaskInitTime＝2002-12-01-18:10:08.012-GMT，

MessageMode＝”search”，

SearchString＝{Title＝“Octopussy”})

All devices is temporarily stored related between TaskID and the task relevant parameter, and its database is searched for.P finds the content of being asked, and therefore, it sends it back S0 with following message:

ContentInfoResponse(

Sender＝NodeID(P)，Receiver＝NodeID(S0)，TaskID＝abc，

MessageMode＝“found?content”，ContentID＝UUID，

LocationID＝UUID，

ContentDescription＝{Title＝“Octopussy”，Summary＝”…”，

Actor＝”Roger?Moore”，Actor＝”Maud?Adams”，Actor＝”…”，

Genre＝”Action”，Keyword＝”James?Bond”，…，

AspectRatio＝”16:9”，…}，Duration＝2:05h，BitRate＝7

Mbps, [, the more information relevant with content])

Since in ContentInfoRequest (" search ") message addressing " all " receivers, except S2, do not need receiver that this request is responded, unless it finds and ask the content of mating, mention especially because S2 is used as receiver: S2 must respond this request, and no matter whether it maintains required content.S2 needs its database of some time searches, and sends following message to S0 when beginning to search for:

ContentInfoResponse(

Sender＝NodeID(S2)，Receiver＝NodeID(S0)，TaskID＝abc，

MessageMode＝“searching”)

Equipment S2 does not find the content of being asked.Because come addressing S2 as " must respond " receiver in ContentInfoRequest (" search ") message, it sends it back equipment S0 with following message, although in S2, do not find required content:

ContentInfoResponse(

Sender＝NodeID(S2)，Receiver＝NodeID(S0)，TaskID＝abc，

MessageMode＝“found?content”，LocationID＝“none”)

The content that the user may find him to search for before the search of all devices is finished dealing with.Therefore, it can make S0 utilize following message cancellation search to handle:

CancelTaskRequest(

Sender＝NodeID(S0)，Receiver＝all，TaskID＝abc)

After receiving this message, all devices stops its search to be handled.Because come addressing S2 as " must respond " receiver in ContentInfoRequest (" search ") message, it sends it back S0 with following message, to confirm CancelTaskRequest (" search ") request:

CancelTaskResponse(

Sender＝NodeID(S2)，Receiver＝NodeID(S0)，TaskID＝abc)

After having sent CancelTaskResponse message to S0, node P and S2 delete TaskID and relevant parameter from temporary storage.Any equipment that has sent CancelTaskResponse message all carries out identical operations.

The user is satisfied with Search Results, and S0 sends a request message to S1, S2 and S3 now, inquires its place capacity, does not use memory capacity and transfer rate so that find out it.Equipment S1, S2 and S3 respond by its place capacity is notified to S0:

DeviceCapabilitiesInfoRequest(

Sender＝NodeID(S0)，TaskID＝bcd，Receiver＝NodeID(S1))

DeviceCapabilitiesInformation(

Sender＝NodeID(S1)，Receiver＝NodeID(S0)，TaskID＝bcd，

DeviceCapabilityInformation{DeviceType＝stationary

storage?device，DeviceServices＝record?or?playback，

MaxCapacity＝100GB，FreeCapacity＝5GB，MaxTransferRate＝30

Mbps，FreeTransferRate＝20Mbps，MaxStreams＝2[，

ActiveStreams＝1，Until＝20:15:00:00]})

DeviceCapabilitiesInfoRequest(

Sender＝NodeID(S0)，Receiver＝NodeID(S2)，TaskID＝cde)

DeviceCapabilitiesInformation(

Sender＝NodeID(S2)，Receiver＝NodeID(S0)，TaskID＝cde，

DeviceCapabilityInformation{DeviceType＝stationary

storage?device，DeviceServices＝record?or?playback，

MaxCapacity＝50GB，FreeCapacity＝40GB，MaxTransferRate＝30

Mbps，FreeTransferRate＝30Mbps，MaxStreams＝2})

DeviceCapabilitiesInfoRequest(

Sender＝NodeID(S0)，Receiver＝NodeID(S3)，TaskID＝def)

DeviceCapabilitiesInformation(

Sender＝NodeID(S3)，Receiver＝NodeID(S0)，TaskID＝def，

DeviceCapabilityInformation{DeviceType＝stationary

storage?device，DeviceServices＝record?or?playback，

MaxCapacity＝300GB，FreeCapacity＝200GB，

MaxTransferRate＝40Mbps，FreeTransferRate＝40Mbps，

MaxStreams＝2})

Perhaps, S0 also can send to RequestDeviceCapability message three all nodes, and is as follows:

DeviceCapabilitiesInfoRequest(

Sender＝NodeID(S0)，Receiver＝NodeID(S1)，

Receiver＝NodeID(S2)，Receiver＝NodeID(S3)，TaskID＝bcd).)

S0 assessment S1, S2 and S3 do not use capacity and transfer rate.S1 does not have the sufficient memory capacity of not using, and S3 provides maximum capacity.Use the memory capacity of the fixed memory device in the network, S0 for balance and select S3 to write down content automatically, and need not to carry out alternately, and ask S3 and P to carry out transmission with the user from P.In the variant of this situation, will omit a Receiver, and the following beginning of message:

InitiateTransferRequest(

Sender＝NodeID(P)，Receiver＝NodeID(S3)，TaskID＝fgh，…)

(variation“B”：Destination＝NodeID(P)，

Source＝NodeID(S3)).

In this case, but only work as the resource requirement time spent, just allow node P to initiate this InitiateTransferRequest:

InitiateTransferRequest(

Sender＝NodeID(S0)，Receiver＝NodeID(S3)，

Receiver＝NodeID(P)，TaskID＝fgh，

TransferPurpose＝”Record”，Destination＝NodeID(S3)，

Source＝NodeID(P)，ContentID＝UUID，LocationID＝UUID，

ContentDescription＝{Title＝“Octopussy”}，Duration＝2:05

h，[Start＝00:00:00:00，End＝02:05:00:00，]

RequestedBitRate＝7Mbps，UseKey＝Key(John’s?James?Bond

friends))

This message request should send the content that is positioned at node P to node S3, and is recorded in node S3 place.ContentID is the UUID that has specified the position of content on node P.TaskID is UUID, and for example can define according to the position of the NodeID of related equipment, the content that will transmit and the time of the task of initiation.If according to its FreeTransferRate, equipment P and/or S3 are too busy this moment, and it will send InitiateTransferResponse (" refusal ") message to S0; So S0 will be by sending CancelTaskRequest message to P and S3, and by P and S3 by CancelTaskResponse acknowledgement messaging S0, cancel this task; Perhaps retry record after a while, or, utilize the After parameter to arrange record according to the Until that from DeviceCapabilitiesInformation message, obtains.

After receiving above-mentioned message, S3 and P confirm request and distribute corresponding resource.The user wants the lineup that will the access authorization of copy of content be given by its defined label " John ' sJames Bond friends " management, and correspondingly indicates S0:

InitiateTransferResponse(

Sender＝NodeID(S3)，Receiver＝NodeID(S0)，

Receiver＝NodeID(P)，TaskID＝fgh，

MessageMode＝”confirmed”，TransferPurpose＝”Record”，

Destination＝NodeID(S3)，Source＝NodeID(P)，ContentID＝UUID，

LocationID＝UUID，

ContentDescription＝{Title＝“Octopussy”}，

[Start＝00:00:00:00，End＝02:05:00:00，]ReservedBitRate＝7

Mbps，UseKey＝Key(John’s?James?Bond?friends))

InitiateTransferResponse(

Sender＝NodeID(P)，Receiver＝NodeID(S0)，

Receiver＝NodeID(S3)，TaskID＝fgh，

MessageMode＝”confirmed”，TransferPurpose＝”Record”，

Destination＝NodeID(S3)，Source＝NodeID(P)，ContentID＝UUID，

LocationID＝UUID，

ContentDescription＝{Title＝“Octopussy”}，

[Start＝00:00:00:00，End＝02:05:00:00，]ReservedBitRate＝7

Mbps，UseKey＝Key(John’s?James?Bond?friends))

Because the value of TransferPurpose parameter is " Record ", Destination node S3 transmits control data and handles: so (or after a while, according to the After parameter), S3 request P sends the content corresponding data to it:

ForwardDataRequest(

Sender＝NodeID(S3)，Receiver＝NodeID(P)，TaskID＝fgh，

ContentID＝UUID，LocationID＝UUID，

[ContentDescription＝{Title＝“Octopussy”}，]

[Start＝00:00:00:00，End＝02:05:00:00])

Equipment P receives the request from S3, and with the content of being asked, sends following response message to S3, thereby begin to transmit content-data from P to S3:

ForwardDataResponse(

Sender＝NodeID(P)，Receiver＝NodeID(S3)，TaskID＝fgh，

ContentID＝UUID，LocationID＝UUID，

[ContentDescription＝{Title＝“Octopussy”}，]

[Start＝00:00:00:00，End＝02:05:00:00，]BitRate＝7Mbps，

Content)

S3 now with recording processing begin to be notified to S0, thereby can be notified to the user:

TransferStatusInformation(

Sender＝NodeID(S3)，Receiver＝NodeID(S0)，TaskID＝fgh，

MessageMode＝“starting”[，TransferPurpose＝”Record”，

Destination＝NodeID(S3)，Source＝NodeID(P)，ContentID＝UUID，

LocationID＝UUID，

ContentDescription＝{Title＝“Octopussy”}][，

Start＝00:00:00:00，End＝02:05:00:00][，BitRate＝7Mbps])

Because S3 control transmits (it being begun by ForwardDataRequest message), S3 sends TransferStatusInformation (" beginning ") message to S0.When the P end data transmitted, it sent following informational message to S3, thereby confirmed to have transmitted complete data.If do not receive this message, S3 will use this true conduct indication, show that for some reason transmission is incomplete, for example owing to force off device etc.:

TransferStatusInformation(

Sender＝NodeID(P)，Receiver＝NodeID(S3)，TaskID＝fgh，

MessageMode＝“end?of?data”，ContentID＝UUID，

LocationID＝UUID[，

ContentDescription＝{Title＝“Octopussy”}][，

Start＝00:00:00:00，End＝02:05:00:00])

S3 finishes record, and the informational message that completes successfully about record below S0 sends, thereby can notify the user:

TransferStatusInformation(

Sender＝NodeID(S3)，Receiver＝NodeID(S0)，TaskID＝fgh，

MessageMode＝“completed”[，TransferPurpose＝”Record”，

Destination＝NodeID(S3)，Source＝NodeID(P)，ContentID＝UUID，

LocationID＝UUID，

ContentDescription＝{Title＝“Octopussy”}][，

Start＝00:00:00:00，End＝02:05:00:00][，Duration＝02:05h，

BitRate＝7Mbps][，StorageSpace＝6.11GB])

Equipment P and S3 remove the distribution to its resource, and S0 notifies user's transmission task to complete successfully now.

The present invention can be applied to wherein may to clash or bottleneck and all the networking fields that should limit conflict or bottleneck.Example comprises the network based on peer to peer technology, as owner zone or UPnP (UPnP) technology.

Claims (13)

1. one kind is the method for the data transmission distribution priority in the network, and described data transmit and comprise:

-first node sends first request, and described first request comprises the indication of the data cell of specific data cell or particular type, and described indication is with reference to the mark related with one or more data cells;

-at least one Section Point receives and analyzes described first request;

-described Section Point detects it can provide the data cell of being asked, and it can provide first message of the data cell of being asked to send indication to described first node;

-described first node receives and selects described first message;

-described first node sends second request to described Section Point at least, request transmits described specific data cell, it is characterized in that described first node is described first request and/or the described second request distribution marker, described identifier is corresponding to first priority;

-described Section Point assessment described the identifier corresponding with described first priority, and according to described identifier, calculate second priority, second priority that is wherein calculated comprises ground floor and second layer part priority, ground floor part priority depends on the data transmission type of being asked and is defined automatically, and second layer part priority is the definition of user or application program, and the type that the data of wherein being asked transmit comprises record, playback at least, stream and non real-time transmit in real time; And

-described Section Point is receiving described second when request, and the described specific data cell of transmission in first transmits wherein gives described first to transmit second priority assignment that is calculated.

2. method according to claim 1, it is characterized in that assessing and distribute to the corresponding described identifier of priority that request in the network and/or data transmit comprises: ground floor part priority at first relatively, if and ground floor part priority is equal, relatively second layer part priority.

3. method according to claim 1 and 2 is characterized in that further comprising the steps of:

-described first node is distributed to described first request with timestamp; And

-described Section Point is assessed described timestamp, to calculate described second priority.

4. method according to claim 3 is characterized in that described Section Point also carries out following steps:

-receiving described second when request, calculate poor between described timestamp time and current time;

-described difference and predetermined value are compared;

If-described difference is selected first algorithm below predetermined value, otherwise select the second different algorithms; And

-according to selected algorithm, calculate the numerical value of described second priority.

5. method according to claim 1 and 2 is characterized in that further comprising the steps of:

-described Section Point receives and arranges from another node and/or point to another request of another node, and detects the priority of distributing to described another request, and wherein said another request causes another transmission on the described network;

-according to detected priority and the priority that calculated, described Section Point before described another transmits, during or afterwards, begin described first and transmit.

6. method according to claim 5 is characterized in that inadequate resource transmits and described another transmission to carry out described first simultaneously, and described method is further comprising the steps of:

The ground floor part priority of these two transmission of-comparison;

If-the described first ground floor part priority that transmits is higher than the ground floor part priority of described another transmission, if perhaps ground floor part priority equates, and described first second layer part priority that transmits is higher than the second layer part priority of described another transmission, then begins described first and transmits; And

-otherwise, if it is real-time transmission, then postpones described first and transmit, if perhaps it is the non real-time transmission and can uses residual resource, then begin described first and transmit.

7. method according to claim 1 and 2 is characterized in that user or application program can revise described second layer part priority, but can not revise described ground floor part priority.

8. method according to claim 1 and 2 is characterized in that interrupting ongoing transmission.

9. method according to claim 1 and 2, it is characterized in that described Section Point can receive a plurality of first requests, and utilizing a plurality of first message response described requests, the timestamp of each self-corresponding first request carries out rank order to described first message according to it.

10. network node comprises:

-device is used for reception and analyzes first request, and the described first request expression first node is transmit leg and specific data cell;

-device is used to detect the data cell of being asked and can be used for described network node;

-device is used for sending first message that the described network node of expression can provide the data cell of being asked to described first node;

-device is used to receive second request, and the described second request request transmits described specific data cell; It is characterized in that

-device is used to assess first priority that is associated with described first request;

-device, be used for according to described first priority, calculate second priority, described second priority comprises ground floor and second layer part priority, described ground floor part priority depends on the data transmission type and the definition automatically of being asked, and second layer part priority is the definition of user or application program, and the type that the data of wherein being asked transmit comprises record, playback at least, stream and non real-time transmit in real time;

-device is used for described second priority assignment is given the transmission of described specific data cell; And

-device is used for transmitting described specific data cell receiving described second when request.

11. network node according to claim 10, it is characterized in that also comprising: device, be used for the priority that evaluate assign transmits for request and/or data, wherein said assessment comprises: at first compare ground floor part priority, if and ground floor part priority is equal, relatively second layer part priority.

12., it is characterized in that also comprising according to claim 10 or 11 described network nodes:

-device, be used to assess the timestamp that is used to calculate described second priority, wherein said timestamp more early, described second priority is high more;

-device is used for receiving described second when request, calculates poor between described timestamp time and current time;

-device is used for described difference and predetermined value are compared;

-device if be used for described difference below predetermined value, is selected first algorithm, otherwise is selected the second different algorithms; And

-device is used for calculating the numerical value of described second priority according to selected algorithm.

13., it is characterized in that also comprising according to claim 10 or 11 described network nodes:

-device is used to receive the request from user, application program or another network node; And

-device is used for according to described request, revises second priority that is calculated.

Images